TLDR: Researchers at Pacific Northwest National Laboratory (PNNL) and Micron have launched the Crete computer system, a groundbreaking testbed featuring 15 terabytes of active memory. Designed to accelerate AI-driven scientific research, particularly in data-intensive fields like chemistry and molecular biology, Crete’s unique architecture allows for unprecedented real-time data exchange among AI models and scientific applications.
RICHLAND, Wash. – The landscape of scientific computing is set for a significant transformation with the unveiling of the Crete computer system, a collaborative effort between the Department of Energy’s Pacific Northwest National Laboratory (PNNL) and Micron. Launched on August 12, 2025, in Richland, Washington, Crete is a novel testbed specifically engineered to support the burgeoning field of AI-driven scientific research, emphasizing massive memory capacity over raw processing speed.
At the heart of Crete’s innovation is its staggering 15 terabytes of active memory, a capacity equivalent to 240 laptops operating simultaneously. This immense memory is co-located directly with the system processors, a configuration deemed unique within the DOE laboratory complex and the broader high-performance computing (HPC) community.
The system’s defining feature is its hybrid memory architecture. It seamlessly integrates tightly coupled Registered Dual In-line Memory Modules with custom Compute Express Link (CXL) controller boards, both developed by Micron. This sophisticated setup, facilitated by XConn Technologies’ I/O memory switch, enables memory to be shared and accessed concurrently by multiple processors. This capability is unmatched by current DOE or commercial HPC systems, addressing critical memory bottlenecks that often hinder large-scale scientific applications.
Mark Helm of Micron stated, “The Crete system represents a bold leap forward in redefining how memory and compute can work to unlock scientific discovery.” The platform is envisioned as a ‘memory chat room,’ where vast databases, complex molecular modeling algorithms, and advanced AI agents can interact in real time. This allows for instantaneous information exchange, facilitating rapid predictions, analysis of experimental results, and the suggestion of new research pathways.
Andrés Márquez, the AMAIS initiative lead and a PNNL computer scientist, highlighted the system’s purpose: “Building upon the innovative vision of the late computer scientist Jim Gray, who introduced the concept of ‘The Fourth Paradigm: Data-intensive Scientific Discovery,’ Crete aims to showcase the potential of this idea by accelerating large-scale, data-intensive, memory-starved scientific workloads that can benefit from AI.” He further explained, “The memory mechanism enabled by the XConn memory switch allows memory that is sitting outside in the network to be easily accessed by multiple hosts at the same time, passing information without going to the hard drive for large workloads.”
Crete is already being deployed to accelerate critical research in chemistry, materials science, and molecular biology. Its development was funded by the Department of Energy’s Office of Science through the Advanced Scientific Computing Research program. Access to this unique testbed is available to DOE Office of Science national laboratory users through the Advanced Memory to Support Artificial Intelligence for Science (AMAIS) initiative.
Also Read:
- Specialized Storage Solutions Propel Enterprise Generative AI into Production
- SurrealDB Revolutionizes AI Agent Memory with Integrated Real-Time Data Management
Beyond its computational prowess, cybersecurity has also been identified as a key priority in the system’s design and operation.
